Abstract:This paper presents a novel deep learning model based on the transformer architecture to predict the load-deformation behavior of large bored piles in Bangkok subsoil. The model encodes the soil profile and pile features as tokenization input, and generates the load-deformation curve as output. The model also incorporates the previous sequential data of load-deformation curve into the decoder to improve the prediction accuracy. The model also incorporates the previous sequential data of load-deformation curve into the decoder. The model shows a satisfactory accuracy and generalization ability for the load-deformation curve prediction, with a mean absolute error of 5.72% for the test data. The model could also be used for parametric analysis and design optimization of piles under different soil and pile conditions, pile cross section, pile length and type of pile.

What problem does this paper attempt to address?

### What problems does this paper attempt to solve? This paper aims to solve the problem of predicting the load - deformation behavior of large - diameter bored piles in Bangkok subsoil under complex soil conditions. Specifically: 1. **High - cost and low - efficiency field tests**: Traditional in - situ pile load tests (static or dynamic) are expensive (the test cost per pile is between $17,000 and $28,000) and time - consuming. Therefore, it is necessary to develop a new method that can accurately predict the pile load - deformation curve. 2. **Limitations of existing models**: - **Fully - connected neural network (FNN)**: Most existing studies use FNN to predict the load - deformation curve of piles, but these models have over - fitting problems and large prediction errors. - **Ignoring sequence data**: Existing FNN models do not consider sequence data such as soil profiles or load - deformation curves at previous time steps, which limits their ability to predict pile deformation behavior. 3. **Complex soil conditions**: The soil layers in Bangkok's subsoil are complex, containing alternating clay and sand layers with high spatial variability. This complexity requires more advanced computational techniques to accurately simulate the pile response under different loading conditions. ### Solutions To solve the above problems, the author proposes a deep - learning model (CTM) based on the Transformer architecture. The main features of this model include: - **Encoder - decoder structure**: The encoder encodes the soil profile and pile characteristics as input, and the decoder generates the load - deformation curve as output. - **Attention mechanism**: By introducing the attention mechanism, the model can better handle sequence data and improve prediction accuracy. - **Combining historical data**: The model incorporates previous load - deformation curve data into the decoder to improve the prediction results. - **Strong generalization ability**: The model shows good accuracy and generalization ability, with a mean absolute error (MAE) of 5.72%, which is suitable for parameter analysis and design optimization under different soil and pile conditions. In conclusion, this paper proposes an innovative method using the Transformer architecture to predict the load - deformation behavior of bored piles in Bangkok subsoil, thus providing an efficient and low - cost alternative for pile design and construction supervision.

Transformer-Based Deep Learning Model for Bored Pile Load-Deformation Prediction in Bangkok Subsoil

Hybrid Deep Learning Model for Prediction of Monotonic and Cyclic Responses of Sand

Developing an Explainable Artificial Intelligent (XAI) Model for Predicting Pile Driving Vibrations in Bangkok's Subsoil

Efficient machine learning model for settlement prediction of large diameter helical pile in c—Φ soil

Prediction of Pile Running During Installation Using Deep Learning Method

Forecasting the Bearing Capacity of the Driven Piles Using Advanced Machine-Learning Techniques

Model Test on Load Transfer Mechanism of a Static Drill Rooted Nodular Pile

Transformer Based Model for Predicting Rapid Impact Compaction Outcomes: A Case Study of Utapao International Airport

Estimation of ultimate bearing capacity of bored piles using machine learning models

Tunnel Boring Machines (Tbm) Performance Prediction: A Case Study Using Big Data And Deep Learning

Quantitative Analysis and Modification of Dynamic P-Y Curve Model for Offshore Wind Turbines Considering Earthquake History Effect Based on Deep Learning

A comparative study on the most effective machine learning model for blast loading prediction: From GBDT to Transformer

Enhancing load prediction for structures with concrete overlay using transfer learning of time–frequency feature-based deep models

Analytical Method for Prediction of Progressive Deformation Mechanism of Existing Piles Due to Excavation Beneath a Pile-Supported Building

Enhanced multi-task learning models for pile drivability prediction: Leveraging metaheuristic algorithms and statistical evaluation

Real-time prediction of TBM penetration rates using a transformer-based ensemble deep learning model

Base resistance of super-large and long piles in soft soil: performance of artificial neural network model and field implications

Transformer based neural network for daily ground settlement prediction of foundation pit considering spatial correlation

Electrical Load Forecasting using a Novel BI-GRU Encoder Decoder Model

A Machine Learning-Based Method with Integrated Physics Knowledge for Predicting Bearing Capacity of Pile Foundations

World Modeling for Autonomous Wheel Loaders